Library Subscription: Guest
Journal of Automation and Information Sciences

Published 12 issues per year

ISSN Print: 1064-2315

ISSN Online: 2163-9337

SJR: 0.173 SNIP: 0.588 CiteScore™:: 2

Indexed in

Pseudoprojection Estimation Algorithms Based on Approximation of Orthogonal Projection Operation

Volume 52, Issue 3, 2020, pp. 13-32
DOI: 10.1615/JAutomatInfScien.v52.i3.20
Get accessGet access

ABSTRACT

The Kaczmarz algorithm proposed in [1] is one of the most effective and most computationally simple one-step estimation algorithms. In a number of subsequent studies the possibility of acceleration the Kaczmarz algorithm by the use of not one but a series of measurements was examined. Research made in [8, 9] was a push for the development of a new class of algorithms − multistage projection algorithms [10-14], in which during the construction of estimates on the n-th step not only new information is used, as it comes in Kaczmarz algorithm, but also information about number of previous steps n-1, n-2 ... . The number of these steps determines the algorithm's memory. In this case, thanks to a better extrapolation and filtering, in some cases it is possible to reduce significantly the time of identification. The implementation of multistep (S-step) projection algorithm requires the computation of the inverse matrix of observations of the dimension S × S. In [11-14] properties of random pseudoinverse matrix and the projection matrix are set. It helped to determine the rate of convergence of these algorithms and conclude that taking into account information about previous S steps is equivalent in terms of convergence speed to reduction of the dimension N in the original space to S. In this paper we propose and investigate pseudoprojective algorithms that have close to the multistage projection algorithm properties, but are simpler to implement. These algorithms use an approximation of the exact projection operation and are based on the one-step Kaczmarz adaptive algorithm. Estimates of the convergence rate of the proposed procedures are obtained. It is shown that the use of such approximation allows, with a slight decrease in the convergence rate of the algorithms, to simplify significantly their implementation and increase computational stability due to eliminating the rotation operation of the observation matrix.

REFERENCES
  1. Kaczmarz S., Approximate solution of systems of linear equations, Int. J. of Control, 1993, No. 57, 1269-1271, DOI: 10.1080/00207179308934446.

  2. Chadeev V.M., Determination of the dynamic characteristics of objects during their normal operation for self-tuning, Avtomatika i telemekhanika, 1964, 25, No. 9, 1302-1306.

  3. Raibman N.S., Chadeev V.M., Adaptive models in control systems [in Russian], Sovetskoye radio, Moscow, 1966.

  4. Aved'jan A.D., Bestimmung der parameter linearer modelle stationarer und instationarer strecken, Messen, steuern, regeln, 1971, No. 9, 348-350.

  5. Rudenko O.G., Estimation of the rate of convergence of one-step stable identification algorithms, Doklady AN USSR, Seriya A, Fiz-mat. i tekhn. nauki, 1982, No. 1, 64-66.

  6. Liberol B.D., Rudenko O.G., Bezsonov A.A., Convergence study of one-step adaptive identification algorithms, Mezhdunarodnyj nauchno-tekhnjcheskjy zhurnal "Problemy upravlenjya j jnformatjkj", 2018, No. 5, 19-32.

  7. Liberol B.D., Rudenko O.G., Timofeev V.A., Modified Kaczmarz algorithm for estimating the parameters of nonstationary objects, Problemy upravlenjya j jnformatjkj, 1995, No. 4, 81-89.

  8. Lelashvili Sh.G., Application of one iterative method for the analysis of multidimensional automatic systems, Skhemy avtomatjcheskogo upravlenjya, 1965, 19-33.

  9. Lelashvili Sh.G., Some questions of constructing a statistical model of multidimensional objects, Avtomatjcheskoye upravlenjye, 1967, 59-96.

  10. Avedyan E.D., Modified Kaczmarz algorithms for estimating the parameters of linear objects, Avtomatjka j telemekhanjka, 1978, No. 5, 64-72.

  11. Ishchenko L.A., Liberol B.D., Rudenko O.G., Projection algorithms for the identification of linear objects, Doklady AN USSR, Ser. A, 1985, No. 7, 62-64.

  12. Ishchenko L.A., Liberol B.D., Rudenko O.G., Adaptive estimation of the parameters of nonstationary objects, Doklady AN USSR, Ser. A, 1985, No. 12, 70-72.

  13. Ishchenko L.A., Liberol B.D., Rudenko O.G., On the properties of a class of multistep adaptive identification algorithms, Kjbernetjka, 1986, No. 1, 92-96.

  14. Liberol B.D., Rudenko O.G., On the properties of projection algorithms for estimating the parameters of nonstationary objects, Doklady AN USSR, Ser. A, 1990, No. 4, 71-74.

  15. Tsypkin Ya.Z., Adaptation and training in automatic systems [in Russian], Nauka, Moscow, 1968.

  16. Polyak B.T., Comparison of the rate of convergence of one-step and multistep optimization algorithms in the presence of measurement interference, Izv. AN SSSR, Tekhnjcheskaya kjbernetjka, 1974, No. 5, 9-12.

  17. Seber J., Linear regression analysis [Russian translation], Mir, Moscow, 1980.

Begell Digital Portal Begell Digital Library eBooks Journals References & Proceedings Research Collections Prices and Subscription Policies Begell House Contact Us Language English 中文 Русский Português German French Spain